How will AI learn cause and affect ?

We are told somethings as absolute truths by other humans.

When we gather these things and analyze them we reach the conclusions of cause and effect. Let us see a simple example we are told the earth pulls things. The glass breaks when it something hard with a high speed. So then we deduce that the glass will break if it gets pulled freely by earth. Then we eventually understand that the gravity is causing the glass to speed up and the affect of that speed are shattered glass pieces.

We learn all of this over a period of time, So my idea is that if we want the AI to learn this we will have to create a syllabus for AI for a cause and affect learning. Then we will have to do this over a long course of time. Humans learn cause and effect of each thing in many instances of experiences. Coming back to our glass and gravity example. So Then humans see things falling down almost every day which reinforces their belief in the gravity law. Then actual experiences of glass and gravity a couple of times makes it a universal truth for us. Now we add a new truth glass pieces are dangerous if u accidentally put ur foot on them. Which will leave to a deeper cause effect understanding that a broken glass is dangerous for us.

My belief is that AI deep learning artificial neural networks will reach that kind of understanding if they are given data in similar fashion for a diverse number of things each being related to another. This will eventually lead to deeper and deeper understanding over years, leading to new hypothesis the. experiences, experiments, observations confirming them. This will help make non biological AI closer to Human AI .